Parallel defeasible argumentation

Abstract

Implicitly exploitable parallelism for Logic Programming has received ample attention. Defeasible Argutmentation is specially apt for this optimizing technique. Defeasible Logic Programming (DLP), which is based on a defeasible argumentation formalism, could take full advantage of this type of parallel evaluation to improve the computational response of its proof procedure. In a defeasible argumentation formalism, a conclusion q is accepted only when the argument A that supports q becomes a justification. To decide if an argument A is a justification a dialectical analysis is performed. This analysis considers arguments and counter-arguments. DLP extends conventional Logic Programming, capturing common sense reasoning features, and providing a knowledge representation language for defeasible argumentation. Since DLP is an extension of Logic Programming, the different types of parallelism studied for Logic Programming could be applied. We propose new sources of parallelism that can be implicitly exploited in the defeasible argumentation formalism implemented through DLP. Both the argumentation process and the dialectical analysis benefit from exploiting those sources of parallelism.